Report non-targeted metabolomics analysis, clustering and differential abundance analysis

1 Loading and processing of data

---

1.1 Count matrix

First, we load and filter the count matrix.

1.2 Check distribution

Log transformation increases normal distribution of sample measurements for the respective metabolites

2 Dimensionality reduction methods

---

2.1 PCA

PCA

3 Clustering and correlation between sample

---

Next, we perform clustering and correlation analysis, and plot these results in a couple of heatmaps. Most of the samples looked fine, no sample was removed as outlier.

3.1 Distance matrix, followed by clustering

## pdf 
##   2

Heatmap of distances

3.2 Correlation between samples

Heatmap of correlations

## pdf 
##   2

3.3 Hierarchical clustering of the samples

Hierarchical clustering

## pdf 
##   2

4 Differential abundance analysis

---

We performed a differential abundance analysis by parametric testing (two-sided t-test). The mean value of replicates per patient-derived cell line was used.

4.1 Average values of replicates per patient-derived cell line

4.2 Differential abundance analysis using two-sided t-test

And these are the results of our differential abundance analysis. Only genes with a signficant adjusted p-value are shown.

5 Visualization of significantly altered metabolites

5.1 Draw boxplots of significantly altered metabolites

5.2 Vulcano-plot

## pdf 
##   2

## pdf 
##   2